LLM-based Query Expansion Fails for Unfamiliar and Ambiguous Queries

TL;DR

This paper investigates the limitations of large language model-based query expansion, showing it often fails with unfamiliar or ambiguous queries, leading to degraded search performance, and provides a framework for evaluation.

Contribution

It systematically analyzes failure cases of LLM-based query expansion due to knowledge gaps and ambiguity, offering insights and a framework for better evaluation.

Findings

LLM-based QE degrades retrieval when knowledge is insufficient.

Ambiguous queries cause biased and narrow search results.

A new evaluation framework highlights LLM limitations in retrieval tasks.

Abstract

Query expansion (QE) enhances retrieval by incorporating relevant terms, with large language models (LLMs) offering an effective alternative to traditional rule-based and statistical methods. However, LLM-based QE suffers from a fundamental limitation: it often fails to generate relevant knowledge, degrading search performance. Prior studies have focused on hallucination, yet its underlying cause--LLM knowledge deficiencies--remains underexplored. This paper systematically examines two failure cases in LLM-based QE: (1) when the LLM lacks query knowledge, leading to incorrect expansions, and (2) when the query is ambiguous, causing biased refinements that narrow search coverage. We conduct controlled experiments across multiple datasets, evaluating the effects of knowledge and query ambiguity on retrieval performance using sparse and dense retrieval models. Our results reveal that LLM-based QE can significantly degrade the retrieval effectiveness when knowledge in the LLM is insufficient or query ambiguity is high. We introduce a framework for evaluating QE under these conditions, providing insights into the limitations of LLM-based retrieval augmentation.